Suggy : An Automatic Query Refinement Agent

We present the agent Suggy, whose aim is the refinement of queries in an Information Retrieval environment on the Internet. It integrates standard Term Frequency and Document Frequency techniques with heuristics that help in the choice of the most characterizing keywords. An experimentation was carried out to confirm that the ideas underlying Suggy can in fact improve both user satisfaction as regards search results and the efficiency of the refinement process.